Local elections are about morale rather than statistics: impact rather than trends. In three weeks’ time, and a long Friday counting in London, will the world have changed?

Of course, if Ken Livingstone does manage to hold on for a third term, Labour leaders will be mighty relieved, and feel that, maybe, they are not inevitably doomed. Conversely, if Boris Johnson wins, the Tories will celebrate a great victory. Ousting the man who has been Mr London for nearly three decades will rightly be seen as their biggest electoral triumph since 1992.

So Labour will be down, and the Tories up, and claiming that they are on their way to winning the next general election. And gloom can be contagious, as Labour MPs have shown in the past few weeks in their bout of self-induced panic and disarray.

These reactions affect political attitudes. But it will be much harder to draw clear-cut statistical conclusions. The Boris-Ken battle is mainly about the two individuals, rather than their parties, which they hardly ever mention. If Mr Livingstone loses, it will primarily be a verdict on him, and, only secondarily, one on the Brown Government. After all, mayoral elections are intended to be about the personalities of the main candidates. In 2004, Mr Livingstone ran well ahead of Labour’s votes for the assembly elections, and Steve Norris well behind the Tory candidates across the capital.

Outside London, the pattern could be even more opaque. Unlike a general election, when all seats are up, there will be a patchwork on May 1, with just a third of the seats being voted on in the metropolitan and other urban districts (last fought in 2004), but all the local council seats in Wales and in some new shadow unitary authorities now being created, such as Durham, Northumberland and the now divided Cheshire.

Even with these caveats, the trends are unlikely to be definite. Local factors are increasingly important, as voters are not necessarily delivering a national verdict. Incumbent councils of whatever party can make themselves unpopular through their actions, and lose. Similarly, insurgents such as the British National Party may notch up quite a few successes but these are likely to be widely spread, and will reflect local conditions.

The main focus is usually on net changes in seats and in estimated national vote shares: are the Tories at 40 per cent or above, and will Labour fall behind the Liberal Democrats into third place? However, it is misleading to extrapolate from these aggregate figures to speculate about the result of a general election still possibly two years away. The correlation between local and general election figures is very poor.

This is not world-weary “contraryism”. May 1 still matters, not least in the obvious sense about who runs London for the next four years, and the control of dozens of councils. The numbers themselves will be interesting, but not deterministic about what happens later. The main impact will be on how MPs and the media talk about politics. Moods are important, but they can change swiftly. Just ask Gordon Brown about the past nine months.

Times Recommends

• Our attitude to kids shows we need to grow up
• More than Play-Doh to plug a deficit
• Murder in Mexico brings the drug war home